Quantization format
Q6_K_L
Q6_K_L has a nominal rate of — bits per weight. We have no measured files carrying this label yet.
From the file· 526 files measured
Nominal bpw
—
from the block layout
Measured average
6.811
526 files
Range
4.606–9.239
varies by architecture
File sizes
0.11 GiB+
up to 325.38 GiB
Real files
one per model, most downloaded first
| Model | Size● | Effective bpw● | vs nominal● |
|---|---|---|---|
| Qwen3.6-35B-A3B | 29.05 GiB | 6.941 | — |
| gemma-4-26B-A4B-it | 21.46 GiB | 6.944 | — |
| gemma-4-12B-it | 9.76 GiB | 7.012 | — |
| Qwen3.5-4B | 3.69 GiB | 6.797 | — |
| gemma-4-31B-it | 25.21 GiB | 6.924 | — |
| Llama-3.2-1B-Instruct | 1.01 GiB | 7.026 | — |
| gemma-4-E2B-it | 4.25 GiB | 7.128 | — |
| Qwen3-8B | 6.54 GiB | 6.864 | — |
| Qwen3.5-0.8B | 0.70 GiB | 6.897 | — |
| Llama-3.1-8B-Instruct | 6.38 GiB | 6.825 | — |
| Ornith-1.0-35B | 28.22 GiB | 6.994 | — |
| Llama-3.2-3B-Instruct | 2.55 GiB | 6.821 | — |
| ThinkingCap-Qwen3.6-27B | 22.43 GiB | 7.042 | — |
| Qwythos-9B-v2 | 7.76 GiB | 6.903 | — |
| Qwen3.5-122B-A10B | 101.31 GiB | 6.957 | — |
| Qwen3-Coder-Next | 61.41 GiB | 6.621 | — |
| Qwen2.5-32B-Instruct | 25.39 GiB | 6.657 | — |
| Qwen3.5-35B-A3B | 29.05 GiB | 6.941 | — |
| Qwen2.5-Coder-7B-Instruct | 6.07 GiB | 6.847 | — |
| Qwen2.5-7B-Instruct | 6.07 GiB | 6.847 | — |
| Qwen3-0.6B | 0.65 GiB | 7.430 | — |
| Qwen3-VL-2B-Instruct | 1.39 GiB | 5.614 | — |
| Jan-v3-4B-base-instruct | 3.55 GiB | 6.916 | — |
| Qwen2.5-1.5B-Instruct | 1.24 GiB | 6.889 | — |
| Qwen2.5-VL-7B-Instruct | 6.07 GiB | 6.288 | — |
●From the filewhat these mean